With the development of mobile Internet, a generation of social networking applications (e.g., an instant messaging application) ensures that communication between people is becoming fast and diverse. In order to prevent loss of users of the social networking applications, various social networking applications provide a function of recommending users, whom a user may be interested in are recommended to the user. However, a conventional user recommending method mainly depends on personal information disclosed by the user to recommend other users. Thus, the accuracy of the recommendation is limited.